Little Flood Victim or AI Fake?

2024-10-03 little flood victim posted by Pastor David Wayne Handley

This image popped in my feeds and I shared it on FB and I immediately got called on for posting a fake AI generated image. That didn’t sit right with me. So, per usual, I journeyed down a rabbit hole to determine the parentage of said image, so that you don’t have to. You’re welcome…

First stop, I went to TinEye.com, reverse image search site, to see where this image may have come from and if there was a way to determine where it might have first popped up. Nothing, TinEye showed no previous spreading of said image.

So, I popped the image into my photos.app to see if any camera information was still embedded in the image file. Nada. This one was scrubbed clean. I doubled back to the person on FB who originally posed the image. David Wayne Handley, a pastor living in Rockville, Indiana, who didn’t have any further information about the image on his FB profile page (except that yesterday he posted a short video reviewing the first McDonald’s to be built in his area and the first McDonald’s that he’s ever visited. Wait.. what?!). Several hours later he would apparently delete or remove the image from his feed (no doubt getting flack for posting an “AI generated image of human suffering”).

Next, I dropped the image into four different online AI image detector website (listed below). Three of the four websites reported back greater than 70% probability that the image was human made. The fourth reported 65% chance that the image was AI generated.

I found a link to the image on a Reddit message board, but all I found there was a series of unhelpful jabs taken at “boomers” for falling for this obviously fake image. But no one had any proof of origin or history. So that didn’t help.

Finally I zoomed in on the image to look for pixel irregularities, etc., that might help detect the nature of the image. With close inspection, it does appear that the brow of the child was manipulated and there is some problems with the fur of the puppy along the back of the head and lower back that almost makes that part of the image look like it was shot against a green screen. The image has clearly been edited. But without more definitive proof that it was entirely created by AI, I’m going to say that this is a photograph that has been heavily (and not perfectly) edited and it may or may not have anything to do with the recent series of hurricanes and the flooding that resulted.

This whole episode reminds me of another image that popped up in my feeds in the summer of ‘23, an image that appeared to be depicting the bravery and comradery of soldiers in battle:

Nothing wrong with the sentiment expressed by the person posting the image or anyone else agreeing with the idea of our soldiers’ bravery and dedication to one another. The problem is that this image was the work of artist and photographer Mark Hogancamp and the soldiers in the image are part of his art project titled MARWEN using G.I. Joe action figures. Hogancamp’s story and why he photographed WWII inspired stories using the action figures was put into a book and documentary film and eventually a movie starring Steve Carell and directed by Robert Zemeckis. Part of Hogencamp’s story is that photography and creating these WWII scenes was part of his therapy after he was tragically beaten nearly to death by some local guys, and lost a lot of his memory. He was beaten because he admitted to liking to wear women’s shoes one time at a small local bar.

It turns out that the image of the soldiers is not at all what the people posting on social media thought, but at the same time they were popular because they do express and communicate the story and meaning that the photographer intended. But it all came from very different circumstances and pain, and still connected with audiences because those feelings were true, even if the story behind the images were fiction.

In a world where Google and Apple sell phones with the capability built in (and heavily advertised) where one can “clean up images’ by removing people and things, or change the setting from a roadside to the beach, or make it look like you jumped eight-feet instead of one foot, it isn’t helpful to respond to said images with comments like, “It’s Fake, you stupid boomers!” At the same time, it would be helpful if more information about the image were included in the image or image file (something that AI organizations are intending to include with future revisions).

Listen, almost all of the images I post online are edited in someway. I crop the image to reframe what I think is important and usually boost the saturation level to “30” because the human eye loves vivid colors. And if you think that some form of editing is not happening to the images you see (especially online), then you are in error. I am not about misrepresenting things, like editing an image so that it looks like I jumped eight-feet to slam dunk a basketball… which was part of a google android ad… ugh. I just feel like maybe we need to step away from shouting at each other online and try to connect with the sentiment being expressed by the image posted. Doctored shit that are just blatant lies can go to hell. As I noted in my original FB post, it’s sad that we have to consider such things when we’re just trying to share actual human experiences. Tread carefully, my friends.

And because it’s Friday, here’s my Video Friday, the trailer to Welcome to Marwen. Enjoy.